AI audio translator

AI Audio Translator translates audio and video media files into selected languages using AI-powered speech recognition, translation, and voice synthesis. Upload an audio file or video, select a target language, and get a fully translated version.

Note that This module is created with the help of AI agents.

Translated videos are not perfect like audio might not sync properly with video.

Features

• Audio Translation Pipeline: Transcribes audio → translates text → generates speech in target language
• Video Translation with Timing: Preserves original video with translated audio track, maintaining lip-sync timing through intelligent segment processing
• Flexible AI Provider Support: Works with any AI module provider (Currently only tested with Gemini)
• One-Click Translation: "Translate" button appears directly on media entity operations
• Format Support: Audio (MP3, WAV, OGG, FLAC, M4A) and Video (MP4, WebM, MOV, AVI)

Post-Installation

1. Configure AI Providers: Go to /admin/config/ai/providers and set up at least one provider supporting Speech-to-Text, Chat (for translation), and Text-to-Speech
2. Create Language Vocabulary: Create a taxonomy vocabulary (e.g., "Languages") with terms like "Spanish", "Hindi", "Sanskrit"
3. Module Settings: Visit /admin/config/ai/audio-translator and:
   • Select your language vocabulary
   • Optionally customize the translation prompt
   • Optionally override AI providers for specific operations
   • Configure FFmpeg path if not auto-detected (video translation only)
4. Translate Media: Go to any audio/video media entity, click "Translate" in operations, select language, and submit

Configuration Tips:

• For video translation, ensure FFmpeg is installed on your server (see Status Report)
• Test with short audio files first to verify AI provider configuration
• Monitor the queue at /admin/config/ai/audio-translator
• Check watchdog logs for translation errors: /admin/reports/dblog

Additional Requirements

Required Drupal Modules:

• AI (^1.0) : Provides AI provider management and operation types
• Media (core): For media entity management
• Taxonomy (core): For language vocabulary

Required AI Provider Capabilities:

• Speech-to-Text (transcription)
• Chat or Text Generation (translation)
• Text-to-Speech (voice synthesis)

System Requirements:

• PHP 8.1 or higher
• Drupal 10.4+ or 11.x
• FFmpeg (required for video translation only; audio works without it)
  • Ubuntu/Debian: sudo apt-get install ffmpeg
  • DDEV: Add webimage_extra_packages: [ffmpeg] to .ddev/config.yaml
  • Verify: ffmpeg -version

Technical Details

Video Translation Pipeline:

1. Extract audio track from video using FFmpeg
2. Detect speech segments via silence analysis
3. Transcribe each segment with AI speech-to-text
4. Translate segments with duration constraints to maintain timing
5. Generate TTS audio for each translated segment
6. Time-stretch segments to match original duration (prevents desync)
7. Rebuild audio timeline with proper timing
8. Mux translated audio back into original video

File Size Limits:

• Audio: 25 MB maximum
• Video: 500 MB maximum